Fair Allocation of Bandwidth at Edge Servers for Concurrent Federated Learning Processes
Md. Anwar Hossen, Fatema Siddika, Wensheng Zhang · 2025
Cloud servers can handle large amounts of data and many devices but have slow and unstable communication over extended networks. Edge servers work with smaller data but have faster and more stable communication with nearby devices. This led to the client-edge-cloud system. As FL grows, more processes need a set-up with multiple FL servers. Edge servers have limited bandwidth and must be shared between FL servers and clients, limiting how many requests they can handle simultaneously. This work explores concurrent FL processes within a three-tier system, with edge servers between edge devices and FL servers. A challenge in this setup is the limited bandwidth from edge devices to edge servers. Thus, allocating the bandwidth efficiently and fairly to support simultaneous FL processes becomes crucial. We propose a game-theoretic approach to model the bandwidth allocation problem and develop distributed schemes to find an approximate Nash equilibrium of the game. Through rigorous analysis and experimentation, we demonstrate that our schemes efficiently and fairly assign the bandwidth to the FL processes and outperform the baseline scheme where each edge server assigns bandwidth proportionally to the FL servers' requests that it receives. The proposed distributed and centralized schemes have similar performance.